At the UK’s Centre for Doctoral Training in AI for Medical Diagnosis and Care, doctoral students are tasked with revolutionary research projects that could potentially revamp and transform cancer treatment


“Our healthcare system is generating a vast quantity of information but often, it’s assessed in isolation,” says neuroscientist Anna-Grace Linton. “I am interested in the potential of AI-driven systems to analyze, order, and identify patterns embedded in medical notes, scoring systems, pathology test and scan results to give physicians a unified picture of a patient’s health status to be able to make a more precise diagnosis.”

Linton’s latest research project targets Patient Reported Outcome Measures (PROMs), a combination of Likert scales responses and free text comments self-reported by patients to reflect on their respective perceptions and experiences of treatment.

Both generic and disease-specific PROMs are increasingly valuable tools to investigate the factors influencing the quality of life, healthcare utilization and survival rates of cancer patients. It has been found that psychometrically reliable and operationally feasible PROMs can provide deep insights into the effectiveness and impact of various treatment regimens, as well as expose the level of symptom severity and the burden of the disease on patients. As such, Linton is applying AI methodologies, including natural language processing on PROMs datasets and unstructured comments given by patients to create predictive models for long-term outcomes following cancer treatment and a better understanding of the factors that can improve patients’ wellbeing and chances of survival.

“If you are diagnosed with Stage Two or Stage Two B gastric cancer and happen to live in Asia, you are most likely to be treated with surgery followed by adjuvant chemotherapy,” wrote Joe Sims, an Astrophysicist from Liverpool. “In many countries, this is a common practice but what if a subset of patients were able to recover with only surgery and not chemotherapy?” This is a question Sims is trying to answer using a combination of deep learning and spatial analytical techniques on hundreds of tissue samples containing tumors extracted from patients with Stage Two or Stage Two B gastric cancer.

Half of these patients have undergone surgery while the other half have received both surgery and chemotherapy. Sims wants to uncover patients who won’t necessarily benefit from post-surgery chemotherapy and provide a pathological explanation to the result. Both Linton and Sims are the first cohort of PhD students who commenced their studies at the UKRI (UK Research and Innovation) Centre for Doctoral Training (CDT) in AI for Medical Diagnosis and Care in September 2019.

CDT in AI for Medical Diagnosis and Care is supported by the Leeds Institute for Data Analytics (LIDA) and is working in close partnership with the Leeds Teaching Hospitals NHS Trust for its vast repository of data and extensive clinical input from a digital pathology network of nine hospitals, seven universities, and MedTech companies. It’s one of the 16 CDT Centres for doctoral training in AI funded by UKRI that has a special focus on transforming cancer diagnosis and care through the application of advanced technologies.

The announcement for the launch of these 16 CDT Centres took place in February 2019 as the UK government recognized the pressing need to groom new talents and experts when AI is continuously deployed to improve healthcare, tackle climate change, and create new commercial opportunities. These 16 CDT Centres were funded by a £100 million investment from UKRI. CDT in AI for Medical Diagnosis and Care has a pioneering batch of 10 students, and it aims to have an intake of 50 students in the academic year thereafter.

“The UK is a world leader in AI. But we can’t be complacent. We need to ensure there are enough talented and creative people with the skills and knowledge to harness and develop this powerful technology,” says David Hogg, Professor of Artificial Intelligence and Director of the CDT in AI for Medical Diagnosis and Care. The PhD students, coming from diverse backgrounds, will be supervised by one clinical and one computer science faculty from the University of Leeds to harness a strong understanding of both the technology and medicine and can think computationally about a problem and its solutions.

“AI has the potential to make a real difference to patients with chronic diseases including cancer,” Professor Hogg adds. “Early detection becomes critically important as AI-driven tools would identify those at risks before symptoms appear and suggest lifestyle changes that would reduce long-term risks, greatly speed up and increase the reliability of diagnostic services such as radiology and pathology, and help doctors and patients select the most appropriate care pathway based on personal history and clinical need. We need more talents, preferably a network of expertise, to arrive at that future we have envisioned.”